Empirical evidence for the severe but localized impact of AIDS on population structure.

Despite considerable speculation on the demographic impact of AIDS, there has been, until now, little scientific evidence to establish its existence or scale. Because of the widespread implications of these predictions, methods to combine demography and epidemiology to study empirical situations have been an urgent priority. This study derives the extent and mechanisms of demographic impacts of AIDS from routine data (the 1991 census) in a severely affected country, Uganda. Three characteristics are of particular note: first, the emergence of demographic impacts much earlier than previously estimated; second, their localization with negative population growth at parish but not at district or national scales; third, a greater impact on the number of children than previously predicted, due as much to changes in population fertility as mortality. The emergence of demographic impacts at this stage highlights original aspects of the interdependence of HIV infection and demographic growth not previously recorded and the need to target preventive interventions to youth in developing countries.

PIP: This paper derives the extent and mechanisms of demographic impacts of AIDS from 1991 census data for Uganda. Reports from Uganda indicate a wide range of different HIV prevalences according to geographical area. For example, HIV prevalence varies from 20% in some areas to 13% in the Rakai and less than 2% in the Pallisa districts. Within Rakai, prevalence varies by parish from 1% to 40%. Analysis of the data points to the potential severity of AIDS; large sections of the population structure pyramid may be missing and there is evidence of negative population growth at the parish level. The study also shows, however, how localized the impact of AIDS is at this stage and that such impact does not apply generally to population growth in Africa or even at the national or district levels. Estimated deficits in adults by age were consistent with the distribution of reported AIDS cases, but the deficits were far greater among children. The magnitude of deficits among 0-4 year olds is not fully explained by direct AIDS mortality, with the model showing the important role of reduced fertility due to HIV mortality in women. There is a need to target youth in developing countries with HIV/AIDS prevention interventions.

An age- and sex-structured HIV epidemiological model: features and applications.

An important challenge in modelling the human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) epidemic is to use the increasing quantity of disease surveillance data to validate estimates and forecasts. Presented is a novel model for forecasting HIV incidence by age and sex and among sentinel groups for which data are available. This approach permits a closer relationship between forecasting and surveillance activities, and more accurate estimates validated to data. As inputs the model uses an estimate of the HIV prevalence, country demographic data, and a profile of the sexual risk of HIV infection by age, to project HIV incidence, prevalence, number of AIDS cases and population. The following examples of the use of the model are given: forecasting HIV incidence in East Africa, by age, sex, and among pregnant women; 3-5-year forecasts of HIV incidence; modelling mixed risk behaviour HIV epidemics in South-east Asia; demographic indicators; and targeting a preventive vaccine by age group.

PIP: As an increasing quantity of HIV/AIDS surveillance data becomes available, methods need to be developed to combine forecasting and surveillance activities to effectively use such data with an eye to improving the validity of projections and guiding where interventions are targeted. A model is described which uses simple empirical inputs to forecast the incidence and prevalence of HIV infection, number of AIDS cases, and mortality by age, sex, and sentinel group. The model can be used to produce HIV infection curves, based upon past and present data, which are then projected into the short-term future. The approach allows a closer relationship between forecasting and surveillance activities, and more accurate estimates validated to data. The model inputs are an estimate of the HIV prevalence, country demographic data, and a profile of the sexual risk of HIV infection by age. These inputs are then used to project HIV incidence, prevalence, the number of AIDS cases, and population. The following examples of the use of the model are given: forecasting HIV incidence in East Africa, by age, sex, and among pregnant women; 3-5 year forecasts of HIV incidence; modeling mixed risk behavior HIV epidemics in southeast Asia; demographic indicators; and targeting a preventive vaccine by age group.

Human immunodeficiency virus infection dynamics in east Africa deduced from surveillance data.

Knowledge of human immunodeficiency virus type 1 (HIV) incidence patterns in East African HIV epidemics like that in Uganda is fundamental for guiding interventions and forecasting the future course of the pandemic, yet they are difficult to determine from surveillance data. The authors deduce hypotheses of HIV incidence dynamics from birth cohort analyses of Ugandan acquired immunodeficiency syndrome (AIDS) incidence from 1987 to 1992 and from the age and sex distribution of sexually transmitted disease: an age dependency for HIV risk; a period effect of varying HIV incidence growth; and a replenishment of HIV-susceptible populations through demographic renewal. The hypotheses are tested by incorporating them into a model that generates patterns of HIV incidence, prevalence, and AIDS cases that are consistent with empiric data. When applied to Uganda, the modeled HIV incidence is characterized by a short temporal concentration of high incidence, followed by a decline, stabilization, and concentration in younger ages. The ensuing HIV dynamics result in a rapid build-up and subsequent stabilization of prevalence and mortality in years 10 and 13, respectively, after epidemic onset. When this model is used to forecast scenarios from 1980 to 2000, HIV prevalence declines in some populations, which is different from earlier scenarios. The techniques presented provide an empiric basis to better direct interventions, forecast epidemic impacts, and evaluate determinants of changing incidence and prevalence patterns.

PIP: Knowledge of HIV incidence patterns helps to guide interventions and forecast the future course of the HIV/AIDS pandemic. The authors deduce hypotheses of HIV incidence dynamics from birth cohort analyses of Ugandan AIDS incidence during 1987-92 and from the age and sex distribution of sexually transmitted disease. The hypotheses are then tested by incorporating them into a model which generates patterns of HIV incidence, prevalence, and AIDS cases consistent with empirical data. Applied to Uganda, the modeled HIV incidence is characterized by a short temporal concentration of high incidence, followed by a decline, stabilization, and concentration in younger ages. A rapid build-up is then envisaged followed by a stabilization of prevalence and mortality in years 10 and 13, respectively, after epidemic onset. When the model is used to forecast scenarios over the period 1980-2000, HIV prevalence declines in some populations, different from earlier scenarios. The techniques presented in this paper provide an empirical basis upon which to better direct interventions, forecast epidemic impacts, and evaluate the determinants of changing incidence and prevalence patterns.

Predictors of survival in HIV-infected tuberculosis patients.

OBJECTIVE: To ascertain predictors of survival in HIV-infected tuberculosis (TB) patients. DESIGN: Retrospective cohort study. SETTING: New York City public hospital. PATIENTS: Fifty-four consecutive HIV-seropositive patients with newly diagnosed TB and no other AIDS-defining illnesses. MAIN OUTCOME MEASURES: CD4+ T-lymphocyte counts, completion of anti-TB therapy, repeat hospitalizations with TB, and survival. RESULTS: Forty-five (84%) of the 54 patients died a median of 15 months after TB diagnosis (range, 1-80 months), five (9%) were alive after a median of 81 months (range, 75-84 months), and four (7%) were lost to follow-up after a median of 42 months (range, 30-66 months). In univariate analyses, disseminated TB, intrathoracic adenopathy, oral candidiasis and CD4 count depletion were each associated with decreased survival. In a multivariate analysis, CD4 count depletion was the only independent predictor of decreased survival. Repeat hospitalization with TB occurred in 10 out of 15 patients who did not complete anti-TB therapy compared with one out of 21 patients who completed anti-TB therapy (P < 0.001). CONCLUSION: The clinical presentation of TB and CD4 count at TB diagnosis are each predictive of survival in HIV-seropositive TB patients. The CD4 count is the only independent predictor of survival.

Preparation for phase III HIV vaccine efficacy trials: methods for the determination of HIV incidence.

OBJECTIVE: Accurate estimates of HIV incidence that reflect the effect of non-vaccine interventions (education, counselling, condom promotion, and possibly sexually transmitted disease treatment) and that may be provided in a Phase III vaccine efficacy trial, are needed so that vaccine trial population sample sizes can be accurately determined. In order to avoid delays in the implementation of efficacy trials, well characterized cohorts must also be developed and available to participate in such trials. We reviewed the potential study populations, the epidemiologic methods for the determination of HIV incidence (using open cohort, closed cohort, and seroprevalence data methods), and the need for the development of population cohorts in preparation for Phase III HIV vaccine efficacy trials. SETTING: Phase III trials in developed and developing countries. METHODS: Comparison of open and closed cohorts and those using seroprevalence data to estimate HIV incidence. RESULTS: Open and closed cohorts each have disadvantages and advantages. However, the open cohort may be more suitable for determining estimates of HIV incidence that reflect non-vaccine interventions and for the development of a well characterized cohort available to participate in efficacy trials. CONCLUSION: Careful preparation of research infrastructures and population cohorts will help ensure the successful conduct of scientifically and ethically sound HIV vaccine efficacy trials in the future.

The global HIV pandemic.

The paper reviews the distribution, transmission patterns, and likely impacts of the HIV pandemic at the beginning of 1993. As of early 1993, a cumulative total of 611,589 cases of AIDS were reported to the World Health Organization (WHO); however, because of less than complete diagnosis and reporting, the WHO estimates that a total of 2.5 million cases of AIDS had actually occurred. As of early 1993, the WHO estimates that there have been approximately 13 million infections, of which about 1 million have been in children. By the year 2000 the WHO predicts that there will be 30-40 million cumulative infections in the world, of which 90% will be in developing countries and almost half will be among women. The epidemic of HIV infection in the decade of the 1980s will result in an epidemic of AIDS in the 1990s, which will place great social and economic strains on many countries, particularly those in many areas of the developing world.

Heterosexual transmission of human immunodeficiency virus type 1 from transfusion recipients to their sex partners.

Using lookback procedures and other methods, we identified and then prospectively followed human immunodeficiency virus type 1 (HIV-1)-infected transfusion recipients and their sex partners to determine AIDS incidence and risks of heterosexual transmission of HIV-1. At enrollment, 7 of 32 (21.9%) female partners of male recipients were themselves infected with HIV-1, as compared with none of 14 male partners of female recipients (p = 0.08). No additional episodes of transmission were observed. The prevalence of advanced immunodeficiency at enrollment was similar in male and female recipients. Male recipients with advanced immunodeficiency (CD4+ lymphocyte count < or = 0.20 x 10(9)/L or a history of clinical AIDS) at enrollment were more likely to have infected their female partners (odds ratio = 7.9; p = 0.03) than men with neither condition. Similarly, AIDS-free survival, as estimated by the product-limit method, was lower among male transmitters than among male nontransmitters (p = 0.01). Transmission was not associated with frequency of unprotected vaginal intercourse. Our data suggest that HIV-1-infected men who develop immunodeficiency rapidly are more likely to infect their sex partners and that the greater efficiency of male-to-female HIV-1 transmission is not explained by a greater number of sexual contacts or more advanced immunodeficiency in index subjects.

HIV-1 seroconversion in patients with and without genital ulcer disease. A prospective study.

OBJECTIVE: To determine the relative risk for human immunodeficiency virus (HIV-1) seroconversion in patients with and without genital ulcers caused by chancroid, syphilis, and herpes. DESIGN: A prospective cohort study. SETTING: An inner-city, sexually transmitted disease clinic. PATIENTS: Patients seronegative for HIV-1 with and without genital ulcers who were followed for a minimum of 3 months. INTERVENTIONS: Questionnaire to obtain data on demographics, sexual behavior, and illicit drug use; testing for HIV-1 at entry and at a minimum of 3 months after entry; medical examination for the presence or absence of genital ulcer disease. RESULTS: Overall, 758 heterosexual men with no history of injection drug use completed the study; HIV-1 seroconversion occurred in 10 of 344 (2.9%; 95% CI, 1.4% to 5.3%) men with a genital ulcer and in 4 of 414 (1%; CI, 0.2% to 2.5%) without a genital ulcer (relative risk, 3.0; P = 0.05). In a multiple logistic regression analysis, those men with chancroid and a new sexually transmitted disease during follow-up each had about three times the risk for HIV-1 seroconversion (P < or = 0.04). CONCLUSIONS: In this group of heterosexual men, chancroid and repeated acquisition of sexually transmitted diseases appeared to facilitate the sexual transmission of HIV-1.

Insight into the infection dynamics of the AIDS epidemic: a birth cohort analysis of New York City AIDS mortality.

The authors analyzed the secular trends of New York City acquired immunodeficiency syndrome (AIDS) mortality from 1980 through the first quarter of 1991 using birth cohort techniques to provide insight into reasons for temporal changes in growth of the epidemic. By disaggregating AIDS mortality data into composite birth cohorts, the authors determined that the slowing in the growth of the epidemic is a result of a leveling or decline in AIDS deaths in male birth cohorts born before 1950 and a continued growth in younger male and all female birth cohorts. This phenomenon is believed to largely reflect earlier human immunodeficiency virus type 1 (HIV-1) infection patterns associated with age-related risk behaviors; however, to some lesser extent, it could also reflect age-related host factors or therapies that may influence the time from HIV-1 infection to death. The findings support the hypothesis that the early infection dynamics of the epidemic were differentially related to age and sex, which resulted in a diffusion of infection from older to younger cohorts and from males to females over time. The future growth of the epidemic will largely depend upon the infection patterns of younger birth cohorts. This method of analyzing AIDS incidence or mortality data may contribute to a better understanding of earlier patterns of HIV-1 infection within a defined population, which will be useful for targeting prevention efforts and improving AIDS forecasting methods.

Survival in a cohort of human immunodeficiency virus-infected tuberculosis patients in New York City. Implications for the expansion of the AIDS case definition.

BACKGROUND: The occurrence of pulmonary tuberculosis in human immunodeficiency virus (HIV)-infected persons is believed to represent a less severe stage of HIV-related disease with a more favorable prognosis than other acquired immunodeficiency syndrome (AIDS)-defining conditions; therefore, it has been excluded from the AIDS definition established by the Centers for Disease Control (Atlanta, Ga) criteria. METHODS: To determine the prognosis of patients with HIV-related tuberculosis, we assessed the clinical, immunologic, and HIV infection status of a cohort of male subjects aged 20 to 44 years who were hospitalized with tuberculosis but without AIDS in New York City hospitals from 1985 through 1986, and we determined their mortality through May 1991. RESULTS: The 58 patients who agreed to participate were largely (90%) nonwhite and had a high prevalence of pulmonary tuberculosis (90%) and HIV infection (53%). Patients who were HIV seropositive had significantly lower CD4 cell counts (median, 0.136 x 10(9)/L; range, 0.013 x 10(9) to 2.314 x 10(9)/L vs median, 0.765 x 10(9)/L; range, 0.284 x 10(9) to 2.333 x 10(9)/L), and, during the follow-up period, an 83% mortality rate that was 7.5 times higher than the 11% rate in seronegative subjects. Survival analyses revealed that for all HIV-seropositive subjects the probability of death at 30 months was 72% and the median survival was 21 months (95% confidence interval, 15.5 to 26.5 months), while for HIV-seropositive subjects with CD4 cell counts of 0.2 x 10(9)/L or less, the probability of death at 30 months was 92% and the median survival was 15.75 months (95% confidence interval, 14.0 to 17.6 months). CONCLUSION: The prognosis for patients with HIV-related pulmonary tuberculosis is poor, and those with CD4 cell counts of 0.2 x 10(9)/L or less have survival patterns similar to that of patients with AIDS. We believe that these data support the expansion of the AIDS case definition to include persons with both pulmonary tuberculosis and severe HIV-related immunosuppression.

Mortality trends in a cohort of homosexual men in New York City, 1978-1988.

Trends in mortality related to infection by human immunodeficiency virus type 1 (HIV-1) and to other causes were examined from 1978 to 1988 in a cohort of 8,906 homosexual men who participated in studies of hepatitis B virus infection in the late 1970s in New York City. HIV-related mortality rates increased from 1 per 10,000 person-years in 1980 to 181 per 10,000 person-years in 1986, followed by a plateau from 1986 to 1988. The standardized mortality ratio among white men in the cohort was 3.7 (95% confidence interval (Cl) 3.4-3.9) as compared with white men from across the United States. Higher HIV-related mortality rates were associated with a higher number of sexual partners, a history of gonorrhea and/or syphilis, and serologic markers of infection with hepatitis B virus. After adjustment for demographics and sexual behaviors, the relative risk of mortality for Hispanic men as compared with white men was 1.5 (95% Cl 1.1-1.9). This study illustrates the large excess in mortality among homosexual men over the last decade, with the excess accounted for by deaths from HIV-related diseases. The recent plateau in mortality may be due to the effect of new treatments and/or the decline in new HIV-1 infections among homosexual men. The excess in HIV-related mortality among Hispanic homosexual men was not explained by differences in demographics and factors associated with the sexual transmission of HIV-1.

Drug use among New York City prison inmates: a demographic study with temporal trends.

In a sample of over 2,500 medical charts from inmates arrested in 1986 in New York City, the rate of intravenous drug use peaked at 43% in male inmates aged 36 to 40. Younger inmates had progressively lower rates of intravenous drug use. Inmates detained for longer time periods and female inmates had particularly high rates of intravenous drug use. Rates of intravenous drug use were 30% among Hispanic, 26% among White, and 16% among Black inmates. Although overall rates of drug use among all male inmates in 1986 have remained unchanged since 1975, drug use was most prevalent in inmates aged 31 to 40 in 1986, the same criminal "cohort" which in 1975 had the highest rate of drug use. Implications for targeting of AIDS and drug-related counseling among inmates are discussed.

Heterosexual transmission of HIV-1 associated with the use of smokable freebase cocaine (crack).

A study of risk factors for HIV-1 infection was conducted at a sexually transmitted disease clinic in an area of New York City where the cumulative incidence of AIDS in adults through mid-1990 was 9.1 per 1000 of the population and where the use of illicit drugs, including smokable freebase cocaine (crack), is common. The overall seroprevalence among volunteers was 12% (369 out of 3084), with 80% of those who were seropositive reporting risk behavior associated with HIV-1 infection, including male-to-male sexual contact, intravenous drug use and heterosexual contact with an intravenous drug user. The seroprevalence in individuals denying these risks was 3.6% (50 out of 1389) and 4.2% (22 out of 522) in men and women, respectively. Among these individuals, the behaviors significantly associated with infection were use of crack and prostitution in women, and history of syphilis and crack use in men. These results suggest that in areas where the level of HIV-1 infection in heterosexual intravenous drug users is high and the use of crack is common, increased sexual activity (including the exchange of drugs or money for sex) may result in increased heterosexual transmission of HIV-1.

Acquired immunodeficiency syndrome prevention. Knowledge, attitudes, and practices of primary care physicians.

We conducted a telephone survey of a probability sample of 473 internists, family practitioners, general practitioners, and obstetrician-gynecologists in New York City (NY) in 1988 to assess their knowledge, attitudes, and practices with respect to the prevention of the acquired immunodeficiency syndrome (AIDS). Although 71% of the physicians had cared for a patient with AIDS and 90% had been involved in ordering the human immunodeficiency virus antibody test, only about a third of them took appropriate sexual histories of new patients and only about a quarter (28%) counseled new patients about reducing the risk of contracting AIDS. Multivariate analysis revealed that physician knowledge about AIDS prevention was associated with younger age, more positive attitude toward homosexual males and intravenous drug users, confidence that counseling would result in behavioral change among patients, and specialty other than obstetrics-gynecology. Results indicate a need for increased training and education of primary care physicians about AIDS prevention.

Syphilis treatment response in HIV-infected individuals.

The adequacy of treatment for syphilis has routinely been evaluated by the serological response, i.e. the rapid plasma reagin test (RPR). Since the description of AIDS and HIV aspects of both the natural history of syphilis and the response of Treponema pallidum to treatment have come under increased scrutiny. With concurrent epidemics of HIV and syphilis in New York City, a serological case-control study was done to determine whether HIV-infected individuals given treatment for primary or secondary syphilis have a modified serological response. All study participants had primary or secondary syphilis and paired specimens available for testing. Cases were defined as people who were HIV-positive and were compared with controls who were HIV-negative. HIV-infected patients with primary syphilis when compared with HIV-negative controls were less likely to have a fourfold or greater RPR decrease or seroreversion within 6 months of treatment [15 out of 28 versus 153 out of 210; odds ratio = 0.4, P less than 0.05]. Cases and controls with secondary syphilis had similar serological responses after treatment for syphilis. Although this study adds to the growing body of literature which suggests that HIV may alter the RPR response, prospective studies are needed to determine definitively whether HIV alters the serological response to therapy in patients with early syphilis.

A method for estimating HIV transmission rates among female sex partners of male intravenous drug users.

There are few data available on the dynamics of heterosexual human immunodeficiency virus (HIV) transmission among women whose only risk factor is sexual contact with intravenous drug-using males. Unlike models which rely on unconfirmed estimates regarding sexual partner selectivity and contact rates between infected and uninfected populations, data from a survey conducted between November 1988 and February 1989 of 1,842 women in New York City provide empirical values for these variables which the authors use in a straightforward estimation model. The authors estimate that the number of new infections among women aged 15-44 years in New York City during 1988 was between 876 and 1,668 and that the number of women already infected through heterosexual contact with male intravenous drug users is between 5,390 and 10,230 among the 1,844,285 women in this age group. These conclusions indicate that male-to-female transmission in New York City is of substantial magnitude and that prevention efforts among male intravenous drug users and their female sexual partners must be appropriately directed.

HTLV-I-associated myelopathy associated with blood transfusion in the United States: epidemiologic and molecular evidence linking donor and recipient.

Six months after receiving 58 units of blood components, a 65-year-old white man from New York City, with no other risk factors for human T-lymphotropic virus type I (HTLV-I) infection, developed HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP). Investigation of blood donors identified a 25-year-old white Hispanic woman from Florida whose platelets had been given to the patient and who was seropositive for the virus on a serum specimen obtained 2 years after the donation. She was born in Cuba and had had 2 sexual relationships with men who either had been born in or had resided in the Caribbean. Polymerase chain reaction (PCR) studies of peripheral blood mononuclear cells indicated that both donor and recipient were infected with HTLV-I. Molecular studies of a 595-nucleotide sequence in the 5' envelope region of HTLV-I indicated that the viruses from donor and recipient were identical in each of 32 positions in which published HTLV-I sequences demonstrate molecular heterogeneity; the donor and recipient viruses were also identical in 2 additional positions in which they differed from all published sequences. Transfusion-associated HAM/TSP has occurred in the United States, but additional cases should be prevented by screening blood donations for HTLV-I. Molecular studies of HTLV-I may prove useful in defining the genetic heterogeneity of HTLV-I isolates in the United States and in studying transmission of this virus.